• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
Checklists, Monitoring, and Multitasking in Cockpit Operations
 

Checklists, Monitoring, and Multitasking in Cockpit Operations

on

  • 4,226 views

Presentation from NASA Ames on the risks of cockpit multitasking and the use of checklists to help mitigate those risks

Presentation from NASA Ames on the risks of cockpit multitasking and the use of checklists to help mitigate those risks

Statistics

Views

Total Views
4,226
Views on SlideShare
4,181
Embed Views
45

Actions

Likes
3
Downloads
124
Comments
0

4 Embeds 45

http://cld9.blogspot.com 29
http://www.slideshare.net 10
http://cld9.blogspot.ca 5
http://feeds.feedburner.com 1

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Checklists, Monitoring, and Multitasking in Cockpit Operations Checklists, Monitoring, and Multitasking in Cockpit Operations Presentation Transcript

    • Checklists, Monitoring, and Multitasking in Cockpit Operations Key Dismukes NASA Ames Research Center Ben Berman San Jose State University Foundation/NASA Ames Research Center FAA Research Review 15-16 October, 2008
    • Forgetting to Perform Procedural Tasks •  20 August 2008: MD-82 on takeoff from Madrid –  Flaps not in takeoff position –  Takeoff configuration warning did not sound •  Similar accidents occurred in U.S. in August 1988 (B727), August 1987 (MD-82) –  Flaps not set and warning system failed •  27 major airline accidents in U.S. between 1987 and 2001 attributed primarily to crew error –  In 5 the crew forgot to perform a flight-critical task –  Did not catch with the associated checklist
    • The Multitasking Myth: Handling Complexity in Real-World Operations Loukia D. Loukopoulos, Key Dismukes, & Immanuel Barshi •  Chapter 1: Introduction •  Chapter 2: What is multitasking and how it is accomplished? •  Chapter 3: The Ideal: flight operations as depicted in flight operations manuals •  Chapter 4: The Real: flight operations add complexity and variability •  Chapter 5: Analysis of concurrent task demands and crew Responses •  Chapter 6:The research applied •  Glossary •  Index •  Appendices: A. Methods, B. Human Agents, C. Perturbations, D. Errors •  References
    • The Multitasking Myth Compared Cockpit Cognitive Demands with FOMs and Training •  Ideal (FOM): Tasks are linear/sequential, predictable, and controllable •  Real (Jumpseat): Interruptions, concurrent tasks, tasks out of sequence, unanticipated new tasks •  Perturbations create multitasking demands –  People overestimate ability to multitask –  Common error: forgetting/failing to perform task element –  Factor in many accidents •  Cognitive analysis of multitasking & prospective memory situations
    • Chapter Six: The Research Applied •  Reviewing and Revising Procedures –  Setting flaps for takeoff –  The original pre-takeoff procedure –  The new pre-takeoff procedure •  Aviation and beyond –  Improving the effectiveness of checklists and crew monitoring –  Strategic management of concurrent task demands –  Training and personal techniques •  Summary of recommendations –  For organizations –  For individuals •  Concluding thoughts
    • Checklist and Monitoring Study •  Update on progress since last year’s FAA research review •  These two crucial defenses failed in many accidents — Why? •  Method: Jumpseat observations and cognitive analysis of task demands •  First step: Identify types of error and surrounding circumstances
    • Data Collection Recently Completed •  60 flights observed at three airlines –  Large U.S. airline and large international airline with world- wide flights and a regional airline •  Aircraft: –  B737 (29) –  A320 (11) –  EMB (10) –  B757 (7) –  B767 (2) –  B777 (1) •  Pilot flying: Captain, 63%; First officer, 37% •  Pilot making the error: Flying pilot, 50%; Monitoring pilot, 50%
    • Preliminary Results •  Errors defined as deviations from published SOP, regulations or good operating practice •  899 errors observed in 60 flights –  Observations consist of narrative descriptions of error and context •  Narratives entered in database –  Exploring ways to categorize and analyze data
    • Phase of Flight at Time of Error Pretaxi: 171 (19 percent) Taxi-out: 78 (9 percent) Takeoff/Initial Climb: 24 (3 percent) Cruise Climb: 205 (23 percent) Cruise: 74 (8 percent) Descent: 210 (23 percent) Approach (Vectors or Final): 89 (10 percent) Landing: 2 (0 percent) Taxi-in: 28 (3 percent) Shutdown/Parking: 18 (2 percent)
    • Errors Per Flight Monitoring: 6.8 ± 3.9 (range: 1-19) Checklist: 3.2 ± 2.9 (range: 0-14) Primary procedure: 5.0 ± 4.8 (range: 0-21)
    • Checklist Error Types Item omitted or performed incompletely 50 Flow/Check performed as Read/Do 46 Responded to challenge without looking 36 Poor timing of checklist initiation 32 Checklist performed from memory 17 Checklist not called for 13 Total 194
    • Checklist Issues Item(s) Omitted from Checklist (50 instances) •  Common outcome, but several clusters of diverse situations •  Cluster: Checklist item deferred and later forgotten –  Example: Early call for Approach checklist; last two items deferred •  Cluster: Checklist interrupted by external agent/event –  Example: Departure Briefing interrupted. Last item never completed •  Dodhia & Dismukes: Interruptions Create Prospective Memory Tasks (Appl. Cog. Psychol, 2008) –  Individuals fail to encode explicit intention to resume interrupted task –  Absence of cues to prompt remembering to resume •  Cluster: Items overlooked without interruption or deferral –  Normal cues absent? Attention diverted? Source memory confusion?
    • Checklist Issues Performing Flow-then-check Procedure as Read-Do (46 instances) •  Problematic: –  Not all flow items are on checklist –  Defeats purpose of redundant check •  Why? –  Inherently tedious to laboriously check habitual task just performed? –  Reversion to old Read-Do procedure after company changed SOP?
    • Checklist Issues Responding to Checklist Challenge without Visually Inspecting Items (36 instances) •  Example: Captain responded “ON” to APU Bleed challenge, but bleed was actually off –  Conceivably a case of looking without seeing •  Example: First officer did not look up from checklist card to verify items on overhead panel •  Why? –  Perhaps relying on memory of having just set an item –  Undermines independent verification
    • Checklist Issues Checklist Performed Entirely from Memory (17 instances) •  Example: –  Captain performed Approach checklist without pulling out card –  Captain performed After Takeoff checklist late without pulling out card –  First officer pulled out card but ran Before Start Checklist without looking at it •  Why? –  Using card is slow and awkward compared to fluent execution from memory –  Response to time pressure? –  Do checkpilots notice and correct this error?
    • Checklist Issues Poor Timing of Checklist Initiation (32 instances) •  Example: First officer, pilot flying, called for In-Range Checklist at 10,000 feet instead of 18,000 feet –  Prospective memory error •  Example: Captain called for Taxi Checklist when aircraft was approaching runway intersection, causing first officer to go head down
    • Checklist Issues Checklist Not Called For (13 instances) •  Example: First officer omitted “Flaps up, After Takeoff checklist” call –  PM failure while attention occupied with other tasks •  Example: First officer omitted Approach checklist on final approach •  Example: Approaching departure runway, captain did not call for Before Take checklist. First officer self- initiated and captain did not act surprised. –  A norm for some crews?
    • Monitoring Error Types Callout Omitted or late 214 Verification omitted 123 Failure to monitor aircraft at level-out 64 Pilot head-down at critical juncture 5 total 406
    • Monitoring Issues Callout Omitted or Late (214 instances) •  Most frequent: Omission of “1000 feet to go” call –  Prospective memory issue: Must switch attention between monitoring altimeter and other tasks. Lack of cues to prompt timely switch •  Most serious: Omission of callouts required during unstabilized approaches –  Example: Monitoring pilot did not call out “Unstable” when approach remained unstable below 500 feet –  Flying pilot can be too focused on trying stabilize flight path to evaluate whether possible to land safely –  Similar to SouthWest 1455 at Burbank and American 1420 at Little Rock
    • Monitoring Issues Verification Omitted (123 instances) •  Example: Neither pilot reset altimeter climbing through FL180 •  Example: Captain verified flap position by looking at and touching flap handle without looking at flap position indicator during Landing checklist
    • Monitoring Issues Failure to Monitor Aircraft (64 instances) •  Example: Captain began cruise cockpit panel scan early and did not monitor level-off by automation –  Poor workload management –  Automation complacency? •  Crew occupied with weather avoidance did not notice fuel configuration EICAS message
    • Monitoring Issues Pilot Head Down at Crucial Juncture (5 instances) •  Example: Captain called for second engine start shortly before crossing a runway, First officer went head down •  Example: First Officer started reviewing final weight data and inputting MCDU while aircraft moving through crowded ramp area •  Problematic workload management –  Interferes with monitoring –  Can lead to snowballing problems as crew get behind aircraft
    • How Often Were Errors Caught and by Whom? Error trapped (18%); error not trapped (82%) When trapped, trapped by: Captain (39%) First officer (40%) ATC (11%) Flight attendant (1%) Aircraft warning system (1%) Jumpseat observer (7%) Crewmember trapping error: Pilot making error (21%) Other pilot (89%)
    • What are the Major Themes? •  Still analyzing data—impressions only •  899 errors seem a lot –  But thousands of opportunities of error on every flight •  Wide range in error rates/flight –  Some due to flight conditions and observer familiarity with aircraft –  Still substantial variation among crews—standardization issue? •  Unrealistic to expect 100% reliability among human operators –  Especially when switching attention among multiple tasks
    • Major Themes (continued) •  Monitoring and checklist callouts are especially likely to be dropped during high workload –  Lose the error-trapping protection when it is most needed •  Subtle reason why error-trapping functions are the first to go –  Primary procedural errors (e.g., setting flaps) give feedback (e.g., takeoff abort) –  Monitoring & callout errors rarely lead to bad consequences (though safety compromised) –  Without feedback loop, errors increase, though pilots may be unaware of it
    • Major Themes (continued) •  Many errors were inadvertent errors of omission –  Prospective memory research: human brain not well equipped to remember to perform tasks that are interrupted, deferred, or performed out of normal sequence. –  The Multitasking Myth provides a cognitive account of this vulnerability and gives detailed countermeasures •  Some errors of omission were not inadvertent –  Performing checklists from memory, etc. –  Correct procedure goes against the grain for fluent performance of habitual tasks –  Pilots “streamline”, perform tasks quickly and fluently but lose the protection provided by the procedure
    • Major themes “Streamlining” of Checklists •  Does training adequately explain to pilots their vulnerability to streamlining and its danger? •  How rigorous is checking of checklist deviations? –  Deviations are subtle and fleeting. Checkpilots focus on big picture –  Without feedback loops, procedures will be streamlined to be fast and to minimize mental workload •  Do companies write stringent, perhaps idealistic procedures but tacitly condone streamlining? –  If procedures are unrealistic, should be rewritten
    • Ways to Improve Checklist Use and Monitoring (from The Multitasking Myth) •  Beyond engineering considerations, procedures must reflect realities of operating conditions and human information processing •  Loukia’s study found SOPs often idealistic, failing to capture dynamic & complex nature of real-world conditions and task demands –  Conflicts arise among procedural demands, operational demands, and human cognitive capabilities •  Recommend companies periodically analyze SOPs for conflicts and hidden traps –  Start with incident reports –  Create team of experienced pilots –  Consult with human factors experts
    • Ways to Improve One Company’s Overhaul of Normal Procedures •  Taxi checklist produced conflict between: –  Procedural demand: preparing aircraft for departure and –  Operational demands: controlling movement of aircraft— following taxi route—maintaining awareness of airport layout, aircraft position, position of other aircraft—communication •  Shifting attention among multiple tasks was a major factor in rejected takeoffs and runway incursions •  In-flight procedures, e.g.: –  Schedule flows & checklists to avoid conflict with transitions between ATC during climb-out and descent •  Performing tasks out of normal sequence often leads to forgetting task elements –  Deferring flaps for movement on contaminated taxiways not necessary for all aircraft types
    • Ways to Improve Checklist Use and Monitoring (continued) •  Training: –  Don’t stop with telling pilots what to do –  Explain what errors occur and why –  Use real-world scenarios, e.g., snowballing workload in unstabilized approaches –  Why quality of checklist execution erodes unwittingly –  Need for slow, deliberate execution that goes against the grain •  Expand workload management portion of CRM –  Traditionally focuses on distribution of tasks and handling overload –  Add ways pilots can avoid amplifying workload problems with better timing of task initiation –  Explicitly address time pressures and dangers of rushing
    • Ways to Improve Checklist Use and Monitoring (continued) •  Insure company policies & practices do not implicitly reward rushing and risky decision-making •  Company guidance for monitoring is much too vague –  Specify what to monitor and when –  Specify and emphasize wording and timing of callouts •  You get what you check and what you reward –  Include how checklists are run and monitoring performed in line and sim checks –  Reward correct use, not streamlining
    • More Information •  Loukopoulos, L. D., Dismukes, R. K., & Barshi, I. (in press). The Myth of Multitasking: Handling Complexity in Real-world Operations. Burlington, VT: Ashgate. •  Berman, B. A. & Dismukes, R. K. (2006) Pressing the approach: A NASA study of 19 recent accidents yields a new perspective on pilot error. Aviation Safety World, 28-33. •  Dismukes, R. K., Berman, B. A., & Loukopoulos, L. D. (2007). The Limits of expertise: Rethinking pilot error and the causes of airline accidents. Burlington, VT: Ashgate. •  Dodhia, R. D., & Dismukes, R. K. (in press). Interruptions create prospective memory tasks. Applied Cognitive Psychology, available online at http://www3.interscience.wiley.com/journal/117926656/abstract?CRETRY=1&SRETRY=0 •  FlightCognition Lab website: http://human-factors.arc.nasa.gov/ihs/flightcognition/